Integrations
Connect the models you already pay for.
The leaders are here, and so are the fast-moving challengers, the open-model hosts and your own endpoints. One governed path, your own keys.
Providers
Every provider on one governed path, each with its status.
Connect the providers you already pay for with your own keys. The same rules, caps and audit record apply to all of them.
OpenAIGPT models, with your own key
AnthropicClaude models, with your own key
Google GeminiThrough your Google API key
AWS BedrockClaude, Titan, Llama and Mistral through role-based access
xAI GrokWith your own key
DeepSeekWith your own key
QwenWith your own key
PerplexityWith your own key
GroqWith your own key
Together AIOpen models, with your own key
FireworksOpen models, with your own key
CerebrasWith your own key
NVIDIA NIMWith your own key
DatabricksWith your own key
OpenRouterOne key, many models
Azure OpenAIOAuth connection to your Azure tenant
Google Vertex AIGemini works today through your Google key; the Vertex connection is in development
Self-hosted modelsvLLM and Ollama through OpenAI-compatible endpoints, on Business and Enterprise
Connects as a custom endpoint.
Provider names and logos belong to their owners and identify compatibility only. Status reflects the connector as built; confirm plan limits in the app.
Identity, monitoring and data
Beyond models: sign-in, evidence and export.
| Capability | Status |
|---|---|
| Single sign-on and SCIM provisioning, through the identity provider you use | Beta |
| Audit log and SIEM stream | Beta |
| BI export | Available |
| Power BI feed with scoped keys | Beta |
| Public API and webhooks | Beta |
How many providers you can connect depends on your plan. Self-hosted models connect through OpenAI-compatible endpoints on Business and Enterprise plans.
Not on the list?
Any API that speaks the OpenAI format connects as a custom endpoint. Tell us which provider you need when you request a review.